Created attachment 307262 [details] Stack trace after resume On an ASUS ROG Ally X (and possibly Legion Go) and a kernel of 6.11+, if Bluetooth is enabled, hibernation will fail on resume with RIP __list_del_entry_valid_or_report.cold. This is under the function hci_unregister_dev 6.10 appears to work however. See the stack trace below. It is from a tainted 6.11 kernel, but the same issue occurs on 6.12.0 and 6.11.y Bisection between 6.10 and 6.11 made some kernels that could not boot, so I could not finish it. I also tried reverting "Fix not handling hibernation actions" and "Bluetooth: btusb: Don't fail external suspend requests" completely which did not help The following lines were used to produce the trace: ``` sudo chvt 2 echo N | sudo tee /sys/module/printk/parameters/console_suspend echo devices | sudo tee /sys/power/pm_test echo disk | sudo tee /sys/power/state ```
I think it is commit ceac1cb0259de682d78f5c784ef8e0b13022e9d9. The way i bisected the kernel was a bit unclean. So I will try to do a clean revert in a few hours.
Created attachment 307265 [details] Revert patch
Reverting commits ceac1cb0259de682d78f5c784ef8e0b13022e9d9 and d09009bc80d9d0d812b988888c40cd86e52eaf1e seems to fix the issue (later is required to revert former). I will do more testing later. I tried to look through the patch to see what could be causing it but I cannot.